Problem with indices in the Polish language

Question

Although I read some postings about indices on the forum, I still have trouble with making an index in Polish. I want to have bold letters before a group connected with a single letter (like in the first picture, which was taken from another post to demonstrate my goal, also with diacritic letters - Ć, Ś). The entries should be also sorted alphabetically (like Ć comes after C)

How can I achieve that?

Here is the code, which was compiled with LuaLaTeX, and makeindex: https://www.overleaf.com/read/rnfrvvbxhfgr

I added also the second picture, which shows the current result.

The code of index is divided into two files (only a short parts, because I'm not sure that it is all):

1. Preambuła/preambula.tex:

(...)
\usepackage{fontspec}
\usepackage{polski}
(...)
\usepackage[makeindex]{imakeidx} %obsługa indeksów 
\makeindex[columns=2, options=-s names.mst -C utf8 -L polish]

2. Zawartość końcowa/skorowidz.tex:

(...)
\printindex
(...)

3. names.mst:

headings_flag 1
heading_prefix "\\textbf\{"
heading_suffix "}\\nopagebreak\n"
delim_0 ": "
delim_1 ": "
delim_2 ": "
delim_r "\\nohyperpage{-}"

4. Example indexes with short part of text:

(... - from Rozdziały/planimetria.tex)
W dowolnym trójkącie dwusieczne\index{dwusieczna}, symetralne
\index{symetralna}, środkowe\index[names]{Srodkowa@Środkowa},
wysokości\index{wysokość} przecinają się w \myuline{jednym punkcie}.
(...)

Answer 1

You can't do it with makeindex because it cannot sort correctly other languages than English. Change to xindy.

Also polski is not compatible with LuaLaTeX, use babel.

Run with -shell-escape enabled, if you want automatic processing with xindy.

\documentclass{article}
\usepackage[polish]{babel}
\usepackage{fontspec}
\usepackage[xindy]{imakeidx}
\makeindex[columns=2, options=-C utf8 -L polish]
\begin{document}
Some text
\index{a}
\index{al}
\index{c}
\index{ć}
\index{s}
\index{sa}
\index{se}
\index{ś}
\printindex
\end{document}

Answer 2

A workaround, if using xindy is an option, could be that you sort words that start with ć or ś under C and S, using something like the following approach. You can use xxx@xxx as argument of the macro \index where everything in front of @ is the sorting string, while the stuff in front of @ is the string that will be printed:

\documentclass[12pt, a4paper, openany]{book}
\usepackage[makeindex]{imakeidx}
\begin{filecontents}{names.ist}
headings_flag 1
heading_prefix "{\bfseries "
heading_suffix "}\nopagebreak\n"
delim_0 " \dotfill "
delim_1 " \dotfill "
delim_2 " \dotfill "
\end{filecontents}
\makeindex[columns=2, options=-s names]
\begin{document}
Foo
\index{a}
\index{al}
\index{c}
\index{czz@ć}
\index{s}
\index{sa}
\index{se}
\index{szz@ś}
\printindex
\end{document}

Answer 3

My personal advice is to switch to upmendex. Create a file named names.ist with

icu_locale "pl"
headings_flag 1
heading_prefix "{\\bfseries "
heading_suffix "}\\nopagebreak\n"
delim_0 " \\dotfill "
delim_1 " \\dotfill "
delim_2 " \\dotfill "

The key point here is the line icu_locale "pl". Just run

upmendex -s names <your file>.idx

Here is the example in previous answers, slightly modified. I don’t use imakeidx, so I don’t know how to deal with upmendex. This is why I run it by hand.

\documentclass{article}
\usepackage[polish]{babel}
\usepackage[noautomatic]{imakeidx}
\makeindex[columns=2]
\begin{document}
Some text
\index{a}
\index{ś}
\index{ć}
\index{s}
\index{al}
\index{c}
\index{sa}
\index{se}
\printindex
\end{document}

As to the line breaking with single letters, babel now provides this feature. See one-letter word at the end of line .